Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

The product of proliferation disrupter is concentrated at centromeres and required for mitotic chromosome condensation and cell proliferation in Drosophila

Tibor Torok/ Peter D. Harvie, Michael Buratovich,^ and Peter J. Bryant^ Developmental Biology Center, University of California at Irvine, Irvine, California 92717 USA

Homozygosity for a null mutation in the proliferation disrupter {prod) gene of Drosophila causes decreased mitotic index, defects of anaphase chromatid separation, and imperfect chromosome condensation in larval neuroblasts and other proliferating cell populations. The defective condensation is especially obvious near the centromeres- Mutant larvae show slow growth and massive cell death in proliferating cell populations, followed by late larval lethality. Loss of prod function in mitotic clones leads to the arrest of oogenesis in the ovary and defective cuticle formation in imaginal disc derivatives. The prod gene encodes a novel 301-amino-acid protein that is ubiquitously expressed and highly concentrated at the centric heterochromatin of the second and third mitotic chromosomes, as well as at >400 euchromatic loci on polytene chromosomes. We propose that Prod is a nonhistone protein essential for chromosome condensation and that the chromosomal and developmental defects are caused by incomplete centromere condensation in prod mutants. [Key Woids: Mitosis; centromere; heterochromatin; chromosome condensation] Received October 30, 1996; revised version accepted December 5, 1996.

The precise transmission of genetic information is en­ has indispensable functions both in meiosis and mitosis sured by the accurate segregation of chromosomes in mi­ (Baum et al. 1994; Murphy and Karpen 1995). For in­ tosis and meiosis. For accurate segregation to occur, the stance, the Drosophila minichromosome Dpll87 re­ chromosomes must undergo a series of precisely regu­ quires a minimum of 200 kb of heterochromatic se­ lated events, including chromosome condensation, ki- quences for stable inheritance (Murphy and Karpen netochore organization, microtubule attachment, and 1995), and the presence of 1000 kb of heterochromatin is sister chromatid disjunction. The centromere appears to critical for normal meiotic pairing and homolog disjunc­ have a key organizing function in most of the above pro­ tion in Drosophila females (Karpen et al. 1996). The cen­ cesses (for review, see Pluta et al. 1995). tral core of a minichromosome centromere in fission The centromere of higher eukaryotes consists of sev­ yeast also must be surrounded by 50-100 kb of repeated eral hundred kilobases of repetitive heterochromatic se­ sequences for completely normal transmission (Clarke quences associated with a disc-shaped protein structure, 1990). the kinetochore, which is the site of spindle attachment The presence of heterochromatic sequences is also es­ (Willard 1990). The importance of the kinetochore in mi­ sential in mitosis. Derivatives of the Drosophila mini­ totic events has long been documented (for review, see chromosome Dp 1187 that show abnormal meiotic chro­ Sluder 1990), but on the other hand the enormous mosome transmission also show a high level of mitotic amount of repetitive DNA in the centromeric hetero­ nondisjunction in neuroblast squashes (Murphy and chromatin was long considered to be functionless (for Karpen 1995). The Drosophila Dp(3;f)Th minichromo­ review, see Pardue and Hennig 1990). However, recent some is mitotically unstable because it lacks large het­ experiments with Schizosaccharomyces and Drosophila erochromatic blocks around the centromere (Wines and minichromosomes have led to the identification of mini­ Henikoff 1992). Deletion of all flanking repeats from the mal DNA sequences needed for centromere functions central core of fission yeast minimal centromeres also and provided direct evidence that the heterochromatin results in mitotic instability (Allshire 1995). The mere presence of heterochromatic DNA is not Present addresses: ^Institute of Genetics, Biological Research Center of sufficient for centromere functions; it also has to be the Hungarian Academy of Science, H-6701, Szeged, Hungary,- ^School of properly organized into a functional centromere by spe­ Biology, University of Sussex, Falmer, Sussex, UK. ^Corresponding author. cific proteins. For instance, if the assembly of centro­ E-MAIL [email protected]; FAX (714) 824-3571. meric repeats is disrupted by microinjection of anti-cen-

GENES & DEVELOPMENT 11:213-225 © 1997 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/97 $5.00 213 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Torok et al. tromeric antibodies (ACA) that affect key protein com­ 20-day larval period. Twenty-day prod larvae show a ponents of the heterochromatin, kinetochore morpho­ striking accumulation of apparently undifferentiated tis­ genesis is inhibited in human cells (Bernat et al.l991). sue in the head region, around and inside the imaginal Mutations in the clr4, rikl, and swi6 genes of fission discs. This phenotype was recorded in the original de­ yeast lead to mitotic instability and chromosome loss, scription as neoplastic imaginal disc overgrowth, but we apparently by disrupting the higher order folded struc­ show here that the imaginal discs are smaller than nor­ ture of the centromeric heterochromatin and thereby mal and that the accumulated tissue mass consists of preventing normal centromere function (Allshire et al. dead cell fragments and blood cells. 1995). prod mutants show massive cell death in proliferating The condensation of interphase chromatin into mi­ cell populations totic chromosomes also depends on the function of non­ histone DNA-binding proteins (Earnshaw and Mackay Homozygous prod larvae have a slower growth rate than 1994). Chromosome condensation mutants also com­ normal and do not reach full size until ~4 days after their monly display other mitotic abnormalities, indicating heterozygous siblings pupariate. At 9 days after egg lay­ that condensation is a prerequisite for subsequent mi­ ing (AEL) the mutant larvae still have tiny, slow-growing totic events. For example, cut mutants of Schizosaccha- and abnormal imaginal discs (Fig. 1, cf. A-D and E-H), romyces pombe are defective in chromosome condensa­ with the eye-antenna discs always somewhat larger than tion and also show abnormal sister chromatid separation the others. Despite the extended larval period the discs during mitosis (Saka et al. 1994). Chromosome segrega­ seldom reach normal size, and they completely lose their tion is blocked in Saccharomyces cerevisiae mutants de­ epithelial organization in the aging mutant larvae. Elec­ fective for topoisomerase II (DiNardo et al. 1984), which tron micrographs show massive cell death in the imagi­ is required for chromosome condensation (Wood and nal discs oiprod homozygotes (Fig. 2B,C). Dead cells and Earnshaw 1990). Sensitive alleles of Responder of Segre­ cell fragments accumulate at the basal side of the disc gation Distorter {Rsp^) in Drosophila cause both aber­ epithelium, and by 9 days AEL blood cells have accumu­ rant chromosome condensation and abnormal meiotic lated around the discs, probably as a response to cell segregation during spermiogenesis (Wu et al. 1988). death. Current experimental data on mitotic and meiotic The number of free-floating blood cells drops dramati­ chromosome behavior clearly support the view that het­ cally in the aging larvae (data not shown), indicating that erochromatin condensation, sister chromatid cohesion, most blood cells are engaged in phagocytosis around the homologous pairing, kinetochore organization, and imaginal discs. Another reason for the low number of proper chromosome segregation are all related, at least in blood cells in mutant larvae is the disintegration of the part, because all of these events depend on the functions lymph glands, which are the source of blood cells (Shres- of heterochromatin-associated nonhistone proteins. The tha and Gateff 1982). Around 9 days AEL the first pair of subject of the present paper is proliferation disrupter lymph gland lobes melanizes (Fig. 1I,J). Melanization [prod], a gene in Drosophila encoding a nonhistone chro­ spreads posteriorly with time, and eventually all of the mosomal protein that accumulates at the centromeric lobes are melanized and start to disintegrate. Large heterochromatin especially of the major autosomes. We lymph gland pieces can often be observed in circulation, report the cloning, molecular and genetic characteriza­ sometimes nearly blocking the dorsal vessel. Electron tion of the prod gene, document the association of its micrographs of mutant lymph glands show that this tis­ protein product with centromeric regions as well as eu- sue also undergoes intense cell death (Fig. 2, cf. E and D). chromatic sites, and show that the product plays an es­ Differentiated plasmatocytes appear to engulf their dy­ sential role in mitotic chromosome condensation and ing neighbors in the mutant lymph gland and eventually segregation. die themselves, only to be engulfed by others. The mutant larval brain is slow-growing and smaller Results than normal (Fig. IJ). By the end of larval life it has ac­ quired an abnormal shape (Fig. IK) and contributes to the Isolation and characterization of the prod mutant amorphous tissue mass in the head. At 20 days AEL prod prod was identified in a systematic P-transposon muta­ larvae show a general disintegration of the internal struc­ genesis screen for mutations with overgrowth pheno- tures. The larvae have an overall transparent appearance types (Torok et al. 1993) and was originally called 1(2)88/ because of severe reduction in the fat body, and the head 10. The prod mutation is caused by a single revertible region is filled with a loose mass of tissue mainly con­ PlacW insertion in polytene chromosome band 56A. A sisting of structureless, fused imaginal discs and the ab­ genetic indication that the insertion results in the null normally shaped brain (Fig. IK). They die at about this condition for this gene came from crossing prod flies to time without pupariating. flies carrying the deficiency Df(2R)P34 (Doane and Du- In summary, all actively proliferating imaginal tissues mapias 1987), which removes the 55E2-56B2,C1 region. appear to be severely affected in the mutant larvae; they The resulting hemizygous larvae showed a developmen­ all show slow growth and eventually degenerate by mas­ tal and chromosome phenotype (see below) indistin­ sive cell death. The penetrance of the mutant phenotype guishable from that of the prod homozygote. Homozy- is 100%, but its expressivity is variable. The degree of gotes of prod die as third-instar larvae after a prolonged melanization as well as the amount of cell death differs

214 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Molecular and genetic characterization of prod

Figure 1. Imaginal discs from wild-type larvae 5 days AEL [A-D] and homozygous prod third-instar larvae 9 days AEL [E-H). {A,E) Pair of first leg discs; {B,F) second leg disc; {C,G] wing, haltere, and third leg disc; {D,H] pair of eye-antenna discs. (/,/) Brain and lymph glands from wild-type 5 days AEL (7) and homozygous piod third-instar larvae 9 days AEL (/). In /, the first pair of lymph-gland lobes (Ig) is starting to mela- nize. [K] Part of the tissue mass dissected from a 20 day prod larval head consisting primarily of distorted brain, fused struc­ tureless leg discs, and eye-antenna discs connected to the mouthhook (m). Un­ stained whole mounts are shown. Scale bar, 0.5 mm.

between individuals even at comparable developmental may be a result of cell death within the clone. The sur­ stages and appears to depend on culture conditions, in­ vival of smaller clones and occasional larger ones may cluding the level of crowding. reflect perdurance of the RNA or protein gene product through several cell divisions after the mitotic recombi­ nation event that initiated the clone. Clonal loss of prod function interferes with oogenesis Germ-line mosaic experiments indicate that the Prod and cuticle development protein is also required for oogenesis. When homozygous Genetic mosaic experiments were performed to test the germ-line clones were induced in heterozygous ovoDl/ requirement for the prod gene product in imaginal disc prod females using the dominant female sterile tech­ development as well as in oogenesis. After irradiating nique (Perrimon and Gans 1983), all 535 irradiated fe­ prod/+ larvae, we failed to recover homozygous mutant males failed to lay eggs. According to the control experi­ clones on the adults. When prod homozygous clones ment, -46 of the 535 females should have contained prod were induced by irradiating heterozygous prod/Minute homozygous germ-line clones. This suggests that oogen­ larvae to provide a growth-rate advantage to the clones esis is arrested in the absence of the prod gene product. (Morata and Ripoll 1975), marked mutant clones were recovered on the resulting adults but they were smaller Molecular characterization of the prod gene and occurred at a much lower frequency than in +/Minute controls. In the few larger clones that were A 6-kb genomic region adjacent to the P insertion was recovered on the legs, the cuticle showed a wrinkled sur­ obtained via plasmid rescue (Wilson et al. 1989), and the face and bristles were defective (Fig. 3). The low fre­ DNA was used as probe to screen a genomic library. quency and small size of homozygous prod clones de­ Finally a 20-kb genomic region around the P insertion spite their growth-rate advantage indicates that the gene was cloned and restriction mapped (Fig. 4A). Genomic product is necessary for cell proliferation and/or viabil­ DNA fragments close to the insertion were then used as ity in imaginal discs and that its function is cell-autono­ probes to screen cDNA libraries. A 1563-bp cDNA was mous. The cuticle abnormalities on the larger clones isolated from an early embryonic library, and a 1412-bp

GENES & DEVELOPMENT 215 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Torok et al.

Figure 2. Transmission electron micro­ graphs. [A] Wild-type eye-antenna disc epi­ thelium, [a] Apical disc surface with mi­ crovilli; [b] basal surface. {B,C] prod eye- antenna disc 14 days AEL with the same orientation as A. Note electron-dense dead cells that tend to accumulate basally. In C, the mutant disc is entirely filled inside with dead cell fragments. (D) Wild-type lymph gland with tightly packed, uni­ formly sized hemocytes. (£) prod lymph gland 14 days AEL. The arrowhead shows the pseudopod of a plasmatocyte engulfing other dying blood cells. Note swollen plas- matocytes full of dead cell fragments. Scale bar, 10 jam.

cDNA from a third-instar larval library. Both of these protein product is acidic with a pi of 5.05, and the se­ cDNAs and the corresponding genomic DNA fragments quence shows no significant matches in current se­ were sequenced. The 1412-bp cDNA turned out to be a 5' quence databases. The VlacW element is inserted in the truncated copy of the 1563-bp version. The 1563-bp 5'-untranslated region of the transcription unit and is cDNA contains a single large open reading frame (ORF) encoding a novel 301-amino acid protein (Fig. 4B,C). The translational start site is probably at the second methi­ onine in this ORF, as this site shows higher similarity to genomic map

the Diosophila consensus translation initiation codon -5-4-3-2-10123456789 (Cavener 1987) than does the first one. The predicted s, B ESE H H BH • ' " • • I M

BSPGXH -E S S <4-—' ' • B

0 75 96 309 376 H—I

"PfacV^

prod protein sequence

MNGKMDRRKK RTSSEQYQMY IDMMESDPIF ATGRVPRDYD LNYLTKKWKE LSDRLNKCSS GPTLTPEEWR KRLNDWKNTT RCKYRRSLLS TEKDISMTSV ETRALDLFGK VPTTGGETML NLKSEKDEHD DEMEELGQRT SVAFQKELQA AVEEAINDEV DEEEMVEEHV DHEDMMEENL AETGITASTT AVNTGGGTYR TIWDNTSFE HVEEDPQTVQ PHAVEYVTSR RPAAPVINPG TASSGNKLIN GELPVKRMRT QPREQIIYEV KNAPRCISNM QAVPPLHSTK LEREPSSLTSR

Figure 4. (^4) Restriction map of the genomic region around the prod insertion. (B) BamHl; (E) £coRI; (G) Bgill; (H) Hindlll; (P) Pstl; (S) Sail, (X) Xhol. [B] Enlargement from the genomic map Figure 3. [A] Homozygous prod mosaic clone on femur that showing the prod coding region from the transcriptional start to extends to the joint. The arrowhead shows a wrinkled cuticle the stop site. PlacW is inserted in the 5' untranslated sequence. surface where few bristles could be formed. The arrow shows The solid region represents the 301-amino-acid coding region; one of the weak forked bristles marking the clone. [B] Wild-type open region represents intron sequences. (B) BamHI; (E) EcoRI. control leg with similar orientation as A. (C) Predicted amino acid sequence of the Prod protein.

216 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Molecular and genetic characterization of prod thereby predicted to cause an early transcriptional stop weakly viable as a homozygote with a Minute-like phe­ (Fig. 4B). The gene contains a single 67-bp intron. notype. Thus, the carboxy-terminal half of the Prod pro­ Developmental Northern blots, containing wild-type tein must be sufficient for almost normal function. polyjA)"" RNA from embryonic, larval, pupal, and adult stages were probed with the isolated cDNAs. The results Transformation rescue of the prod mutant (not shown) revealed 1.6-kb and 1.4-kb transcripts at all tested stages. The 1.6-kb transcript is expressed highly The 8-kb Xiiol-Hindlll genomic DNA piece containing throughout development, whereas the 1.4-kb transcript the coding region and a 4.5-kb 5' DNA piece with pre­ is expressed very weakly in all stages except early em­ sumed regulatory sequences (Fig. 4A) was cloned into a bryos where it is not detectable. Transformation rescue P-transformation vector and transformed into flies. experiments using cDNA constructs (see below) show When chromosomes containing this insert were trans­ that the 1563-bp cDNA is sufficient to perform the prod ferred into a homozygous prod mutant background, the function. mutant phenotype and lethality were fully rescued, con­ Northern blots with RNA samples from homozygous firming that the entire prod function is contained in this mutant larvae and wild-type controls were probed with fragment. genomic DNA segments adjacent to the insertion and We also tried to rescue the mutant phenotype by ex­ with the isolated cDNAs. The results revealed that the pressing the prod cDNA with the Gal4/UAS (upstream 1.6-kb transcript and its 1.4-kb shorter version are miss­ activating sequence) system (Fischer et al. 1988) in a ho­ ing from the mutant (Fig. 5A, lane 3), confirming that the mozygous prod mutant background. When UAS-cDNA observed phenotype is the null condition for this gene. constructs are crossed to different Gal4-expressing lines, We remobilized the FlacW transposon of prod to re­ the Gal4 protein activates transcription of the cDNA cover new alleles and small deficiencies by imprecise according to the Gal4 expression pattern (Fischer et al. excision. Only a low frequency of mobilization was 1988). The 1563-bp cDNA was subcloned into a UAS achieved, but this produced one small deficiency, expression vector and transformed into flies. Antibody Df(2R)prod-K, which removes at least 300 bp from the 5' staining of Western blots from induced UAS-Prod trans- end of the prod coding region. Northern blots (Fig. 5A, formants showed that the protein made from the trans- lane 1) show that the deficiency restores transcription of gene is the same size as the wild-type Prod protein (Fig. the gene by removing the FlacW element but makes a 5B, lane 3), indicating that the entire coding region is truncated version of the transcript. Df(2R)prod-K is present on the transgene. Of the six different Gal4 lines tested, Gal4 69B gave the best partial rescue of the mutant phenotype, which was a shift from larval to pupal or pharate-adult lethal­ B ity. These larvae had remarkably normal imaginal discs before pupariation, with only slightly abnormal folding. Kb 1 2 3 KD 2 3 Gal4 69B is known to induce a high level of expression in 106 - imaginal discs during the larval period (Brand and Perri- 2.37- 69 - mon 1993). 43 - 1.35- i 4. Surprisingly, UAS-prod expression in a wild-type background was lethal at the embryonic or early larval 28 - stages with all of the tested Gal4 lines except hsGal4 18.8 - (heat-shock Gal4). The slight ubiquitous overexpression caused by the hsGal4 construct without heat shock re­ sulted in prolonged development, weak semisterile adults, and rough eyes. A 30-min heat shock at 37°C killed all animals at all tested developmental stages. Figure 5. (A) Northern blots of poly(A)'^ RNA probed with the These experiments suggest that any overexpression of 1563-bp prod cDNA. The filter was overexposed to reveal any transcripts in the mutant. (Lane 1) Homozygous Df(2R)pTod-K this gene is lethal and that its expression must therefore larvae with a truncated transcript; (lane 2) wild type larvae; (lane be regulated strictly and quantitatively. 3) homozygous prod larvae. {Bottom) A control hybridization of the same filter with a p-tubulin probe. {B] Western blot probed The Prod protein is locahzed mainly at the second with affinity-purified Prod antibody. (Lane 1] Homozygous Df(2R)prod-K fly; (lane 2) wild-type fly; (lane 3) transgenic and third autosomal centromeres and at muhiple UASprod; hsGal4 fly after a 30-min heat shock at 37°C. The euchromatic sites multiple bands in lane 3 are presumed to reflect multiple trans­ Polyclonal antisera were prepared from rabbits immu­ lation starts from the transgene. The strongest band that is the nized with a glutathione S-transferase (GST)-Prod fusion same size as the wild-type product in lane 2 probably starts from the second methionine of the largest ORF. This site shows the protein and affinity purified. On Western blots the puri­ closest resemblance to the Drosophila consensus translation fied antibodies clearly recognized one strong protein initiation sequence C/AAAA/C ATG. (Bottom] The same filter band with approximately the expected molecular weight probed with anti discs-large antibody (Woods et al. 1996) as a (34.3 kD) and a very weak smaller band (Fig. 5B, lane 2). control. It also recognized a smaller version of the protein from

GENES & DEVELOPMENT 217 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Torok et al. the homozygous; weakly viable Df(2R)prod-K allele (Fig. set). In prophase (condensing) and telophase (decondens- 5B, lane 1), consistent with the presence of a truncated ing) chromosomes the main localization remains centro­ transcript in these flies. Immunostaining showed that meric, but the extracentromeric staining becomes more the protein is expressed ubiquitously in embryos, larvae, dispersed. In interphase the Prod protein is accumulated and adults. The protein shows a characteristic nuclear at the apical nuclear surface, and it also shows a dis­ localization that is modified during the cell cycle. persed nuclear staining (Fig. 6C,D). The apical nuclear To follow the distribution of Prod protein through the surface staining probably also represents centromeres, cell cycle we examined early gastrulae, in which small because this interphase arrangement agrees well with groups of cells enter mitosis together in mitotic domains the "Rabl orientation" of interphase chromosomes de­ (Foe 1989) and most mitotic phases can be observed in scribed in early Drosophila embryos (Ellison and Howard close proximity to one another. During mitoses the an­ 1981; Foe and Alberts 1985), where the Hoechst 33258- tibody staining is clearly localized to the centromeric positive condensed AT-rich centromeric heterochroma- regions of the second and third chromosomes (Fig. 6A,B). tin is positioned against the apical nuclear envelope. All On metaphase and anaphase chromosomes this localiza­ of the examined proliferating tissues including larval tion is especially obvious, although there is also a imaginal discs, brains, imaginal nests, ovaries, and testes weaker staining along all chromosome arms (Fig. 6B, in­ of both larvae and adults showed a similar nuclear dis-

Figure 6. {A,B] Immunostaining of a mitotic domain on a wild-type stage-9 gastrula, showing most mitotic stages. iC,D] Optical cross section of a cellular blastoderm embryo in interphase, showing Prod-positive centromeres at the apical surface of the nuclear envelope. The arrow points to the apical cell surface. A and C are triple stained with ^-tubulin in green (fluorescein), chromosomes in blue (TO-PRO), and Prod-protein in pink/red (Cy3). B and D show only the Prod signal, demonstrating that the protein is also present on chromosome arms. [A,B (inset)] An enlarged chromosome set, where Prod staining is primarily concentrated on the second and third chromosome centromeres. (£) Immunostaining of the first gonomeric mitoses from a freshly layed egg. Note that the spindle is composed of two units and all 16 centromeres are Prod positive. One chromosome (probably the Y) left the metaphase plate earlier. Staining is the same as in A. (F) Optical cross section of a stage 6 egg chamber from a wild-type adult female ovary stained with Prod antibody. The oocyte (o) shows very strong staining compared to the follicle cells (f) and the polytene nurse cells (n). Scale bars, 10 lam.

218 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Molecular and genetic characterization of prod tribution of the protein during the cell cycle. The Prod important role in mitotic cell division. To test the pos­ staining of interphase and mitotic chromosomes of lar­ sible role of prod in mitosis we made mitotic chromo­ val neuroblast squashes (see below) was identical to that some preparations (Gatti and Goldberg 1991) from mu­ shown for embryos. However, the first mitotic division tant larvae of different ages. These neuroblast squashes, in early embryos seems to be exceptional because all four without colchicine and hypotonic treatment, showed chromosomes show Prod staining (Fig. 6E). This may be several mitotic abnormalities. because there is Prod protein accumulation during game- Most obviously, prod tissue shows a significantly togenesis (Fig. 6F), providing an exceptionally high pro­ lower mitotic index than that seen in wild type. The tein level for the first mitotic division. The high protein number of cells in mitosis per microscopic field in the level in the developing oocyte is also in agreement with mutant at 5 days AEL (0.1) was less than half the number germ-line mosaic data, showing the requirement of prod seen in wild type (0.26). This value decreased in aging for oogenesis. Chromosomes from prod homozygous lar­ larvae, so that in 2-week-old mutant larvae hardly any val brain and imaginal disc cells did not show any de­ mitotic cells could be found. tectable Prod staining (data not shown) The amount of A small proportion (-10%) of the metaphase figures on protein present in the mutant larvae probably dimin­ untreated preparations showed incomplete condensation ishes below the level detectable by immunostaining. at the centromere (Fig. 8B,C) reminiscent of the pheno­ On salivary gland polytene chromosomes the anti- type of some chromosome condensation mutations in Prod antibody reproducibly stains >400 sites including Drosophila (Gatti and Baker 1989). To better examine both bands and interbands. However, the chromocenter the chromosome morphology we made colcemid- and does not stain (Fig. 7A). Nearly all Prod-positive bands hypotonically treated preparations, in which a large are stained strongly also with DAPI (Fig. 7B,C), suggest­ number of cells accumulate in metaphase, chromosomes ing that the protein may preferentially associate with are better spread, and sister chromatids often separate AT-rich regions (Kapuscinski 1995). It probably associ­ (Gatti and Goldberg 1991). On metaphase-arrested prepa­ ates with specific AT-rich structures because not all rations the mutant larvae also showed significantly DAPI-positive bands stain with the Prod antibody. Sali­ fewer mitotic cells than in wild-type and the abnormal vary glands of prod homozygotes are poorly polytenized condensation was more pronounced. Mutant chromo­ and do not show any detectable Prod staining. some arms were about the normal length but often wider than wild-type or irregularly condensed, whereas centro- meric regions were elongated and thinner, apparently be­ cause of undercondensation (Fig. 8J,K). Sometimes the The prod mutation causes chromosome elongated centromeric region was so thin that the two under condensation and failure of the metaphase- chromosome arms seemed to be separated (Fig. 8}, inset). anaphase transition Typically a series of different grades of undercondensa­ The protein localization, the mutant phenotype, and the tion, from normally condensed chromosomes to exten­ mosaic data all suggest that the prod gene product has an sively undercondensed ones, were observed within an

-47D - 47D

-48C -48C

SOD ^„^ / 52C

Figure 7. [A] Salivary gland polytene chromosomes stained with Prod antibody. The arrow points to the chromocenter, which is not significantly stained with Prod. Chromosomes were stained red with propidium iodide; Prod staining (fluorescein) is green in interband and yellow in band regions. [B] Enlarged segment of polytene chromosome arm 2R stained with DAPI. (C) Prod staining of the same polytene segment as B. Nearly all Prod-positive regions are also strongly stained with DAPI. Scale bar, 10 ]am.

GENES & DEVELOPMENT 219 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Tiirok et al.

'^,< we observed randomly arranged chromosomes with separated sister chromatids (Fig. 8F,G) resembling PSCS 4 ^ (premature sister chromatid separation), where sister r chromatids separate precociously, before the onset of anaphase (Miyazaki and Orr-Weaver 1994). The ran­ 1 domly arranged chromosomes sometimes were clearly A I derived from abnormal anaphases, where subsets of chro­ mosomes apparently were not moving toward the poles F G , iH^f^ (Fig. BE). In many of the abnormal figures the scattered chromatids had a fading Giemsa staining, probably be­ cause they had started to disintegrate (Fig. 8H). Appar­ ently in the mutant only a small proportion of meta­ phase cells enter anaphase and finish mitosis. Cells with scattered chromatids, unable to complete anaphase, probably undergo cell death.

Discussion The chromosomal morphology and the abnormal meta­ phase and anaphase frequencies in the prod mutant, to­ gether with the distribution of Prod protein and the de­ velopmental phenotype, suggest that the gene product is Figure 8. Neuroblast squashes from 5 day wild-type {A,D,I) larval brains and 5 day homozygous prod {B,C,E-H,J,K) larval required for normal chromosome condensation, segrega­ brains. Samples A-H were untreated; I-K were treated with col- tion, and cell proliferation. The low proportion of mi­ cemid and hypotonic solution before squashing. [A] Wild-type totic cells due to anaphase failure is consistent with the metaphase with well condensed centromeres. {B,C] prod meta- developmental phenotype, the slow growth of proliferat­ phase chromosomes with poorly condensed centromeres (ar­ ing tissues, and massive cell death. The poorly con­ rowheads). Most sister chromatids in C show abnormal separa­ densed centromeric heterochromatin observed on mu­ tion. (D) Wild-type anaphase. (£) Near normal anaphase from tant centromeres correlates with the absence of Prod pro­ pwd brain, with lagging chromatids. [F-H] Abnormal anaphase- tein from these chromosomes. The poor centromere like figures from prod brains. Chromosomes are fully condensed condensation probably interferes with kinetochore orga­ and sister chromatids have separated, but the chromatids are nization on prod chromosomes, preventing the onset of irregularly positioned. The chromosomes in H are scattered and normal anaphase. It seems likely that the anaphase fail­ fading. [I] colcemid-treated wild-type chromosomes with sister chromatids separated. The arrowhead shows centromeric con­ ure leads to cell cycle arrest and/or cell death. striction. (/) Typical prod chromosome set with wider under- condensed arms and elongated centromeres (arrowhead). Inset shows the same with separated sister chromatids. Compare cen­ Late larval phenotype tromeres on / {inset) with those in /. [K] prod chromosomes with The fact that mutant homozygotes survive to late larval extremely undercondensed centromeres (arrowhead) and chro­ stages suggests either that embryonic and early larval mosome arms. Sister chromatids are separated. Scale bar, 2pm. stages do not require Prod or that the product is required at all stages but that mutant homozygotes are supported through early development by gene product (either RNA individual larva. The proportion of undercondensed or protein) deposited in the egg by the heterozygous chromosomes also showed a great deal of variation be- mother. Work on other Drosophila mitotic mutants has tw^een individuals. In some colcemid-treated prepara­ shown that late larval lethality, together with poorly de­ tions about half of the metaphase chromosomes showed veloped imaginal discs, can be indicative of a defect in signs of undercondensation, whereas in others only a few the cell cycle (Gatti and Goldberg 1991), given maternal showed this effect. This variable expressivity of the mi­ rescue through the earlier stages. We have been unable to totic phenotype is consistent with the similar variability test the idea of maternal rescue directly because germ- observed at the developmental level. line clones of mutant cells fail to make eggs and there­ Another anomaly in the mutant is that fewer ana­ fore do not produce the progeny that would be needed for phases are seen than in wild-type controls. To measure such a test. However, the high protein expression in oo­ the anaphase frequency we compared the number of cytes, the high RNA expression in early embryos, and metaphase figures with the number of anaphase figures the same Prod protein localization during early syncytial within an individual from untreated preparations (Gatti divisions as in later stages suggest that the protein is also and Baker 1989). The metaphase/anaphase ratio was 4.6: required in early stages and is supplied maternally. This 1 (17.9% anaphase) in wild type, compared with 16:1 interpretation implies that the gene product is long-lived (5.9% anaphase) in mutants at 5 days AEL. In making the and persists in the cell over many mitotic cycles to allow above calculation only anaphases that appeared normal embryonic and larval development. The growth of imagi­ or nearly normal were taken into account. In the mutant nal discs and other proliferating tissues in the larva can

220 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Molecular and genetic characterization of prod lead to dilution of the maternal product and therefore In addition to its conspicuous concentration at centro­ would require zygotic expression to prevent the mutant meres. Prod is also present at specific sites along the phenotype from developing, as in other known mitotic chromosome arms in polytene chromosomes and is seen mutants (Glover 1989; Gatti and Baker 1989). The ability more diffusely on the arms of mitotic chromosomes. The of some homozygous prod mutant mitotic clones to sur­ polytene chromosome staining pattern suggests that the vive for a few rounds of cell division also indicates per­ staining pattern seen during diploid interphase may also sistence of the prod gene product. reflect dispersed staining of chromosomes. It is probably Maternal rescue may also help to explain the variabil­ present at the same genomic sites on both mitotic and ity of the mutant phenotype, as the maternal product polytene chromosomes, but the distribution is clearer on may not be equally distributed between daughter cells polytenes because of their larger size. during division, leading cells to die at variable times dur­ The distribution of Prod is consistent with a role for ing development. In some individuals this process can this protein in both centromeric and euchromatin con­ lead to dramatic accumulation of dead cells and cell frag­ densation. Drosophila chromosomes are partially decon- ments in the larval head, the region of most active cell densed by Hoechst 33258 (Gatti et al. 1976), resulting in proliferation, which is further extended by accumulation a chromosomal phenotype similar to that produced by of phagocytic blood cells and cell fragments that drift the prod mutation. Centromere condensation in the from other parts of the body via the circulation. mouse is also incomplete in the presence of Distamycin A, which binds to AT-rich centromeric repeats (Radic et al. 1987). These drugs are thought to act by blocking the Protein localization interaction of condensation-specific proteins with bent We do not yet know whether the Prod protein binds to AT-rich structures (Radic et al. 1987). The similar stain­ DNA directly, or indirectly, through other proteins. It ing pattern of Hoechst 33258 and Prod-positive regions does not show homology to known DNA-binding do­ on mitotic chromosomes, as well as DAPI and Prod-posi­ mains, and its acidic pi would be more consistent with tive bands on polytene chromosomes, raises the possi­ binding to other chromosomal proteins or protein com­ bility that Prod might act in chromosome condensation plexes. In spite of the concentration of Prod on the major by being part of protein complexes associated with some autosomes in most preparations, its function does not AT-rich DNA structures. appear to be restricted to these chromosomes. Both ab­ normal centromere condensation and abnormal sister chromatid separation were also observed on X and Y Centromeric Prod function chromosomes in the mutant (Fig. 8B,C). Preparations The association of Prod protein with the a-heterochro­ from the first embryonic gonomeric mitosis where all matin of major autosome centromeres and the presence centromeres showed Prod staining (Fig. 7E) also suggest of abnormally condensed centromeres in the mutant in­ that Prod may have a general function on all centro­ dicate that the prod gene has an essential function in the meres. The high protein concentration on the two major condensation of centromeric heterochromatin. There is autosomes may represent only a quantitative difference increasing evidence that functional kinetochore forma­ between these chromosomes and the X, Y, and fourth tion requires proper condensation of the centromeric chromosomes, with the smaller amount of protein on heterochromatin. Microinjection of ACA antibodies the latter chromosomes being sufficient for normal func­ causes metaphase arrest by inhibiting kinetochore mor­ tion. phogenesis in human cells (Bernat et al. 1991). ACA an­ The prod gene product at centromeric regions is appar­ tigens (CENP-B in particular) are localized not in the ently not a component of the kinetochore, as the centro­ kinetochore itself but in the surrounding heterochroma­ meric staining does not appear to change during the cell tin. Based on electron microscopy data, Bernat et al. cycle, whereas the kinetochore is known to be disas­ (1991) suggested that ACA antibodies mayx prevent sembled during interphase. It is probably a chromosomal proper condensation of the heterochromatin, 'thereby protein that interacts with centromeric heterochromatin preventing kinetochore organization. Fission yeast mu­ sequences that remain condensed throughout the cell tations in clr4, rikl, and swi6 genes lead to mitotic chro­ cycle (Carmena et al. 1993). The lack of centromeric mosome loss and at the same time alleviate transcrip­ staining in polytene cells also implies an association of tional repression of genes placed adjacent to the centro­ the Prod protein with the a-heterochromatin, which is mere (Allshire et al. 1995). The products of clr4, rikl, and highly underrepresented in polytene chromosomes (Gall swi6 genes presumably influence position-effect varie­ et al. 1971; Miklos and Cotsell 1990). The strong staining gation (PEV) and centromere function by changing the of the second- and third-chromosome centromeres in mi­ structure of the centromeric heterochromatin. totic cells suggests that the protein might associate with We propose that the abnormal anaphases in the prod specific a-satellite sequences that are more abundant on mutants are a direct consequence of incomplete centro­ the heterochromatin of these chromosomes. One of the mere condensation. Proper attachment of spindle micro­ Hoechst-positive AT-rich satellite repeats (AATAA- tubules to the chromosomes requires a well-organized CATAG) is specific for the second and third chromo­ kinetochore, and if spindle attachment is abnormal, the somes as determined by in situ hybridization (Lobe et al. anaphase transition is arrested by a cell-cycle checkpoint 1993) and may provide a preferred Prod-binding site. (Li and Nicklas 1995). The scattering of chromatids in

GENES & DEVELOPMENT 221 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Torok et al. the abnormal prod anaphases indicates that they are ei­ duced viability and rough eyes (Farkas et al. 1994), very ther not attached to the spindle or cannot move along it. similar to the prod weak overexpression phenotype. Trl However, some of the chromatids do seem to move, and is also required for mitoses, as chromosome segregation the Prod protein does not localize to the spindle appara­ and condensation defects were observed in Trl mutants tus, suggesting that it is spindle attachment rather than (Bhat et al. 1996). The GAGA factor activates transcrip­ spindle formation that is impaired in the prod mutant. tion by binding to the promoter regions of many genes Although many of the abnormal prod anaphase-like and helping to keep the chromatin structure accessible chromosomes seem to be properly condensed, it is pos­ to RNA polymerase (O'Brien et al. 1995). sible that undercondensation that is not discernible by These results raise the possibility that the prod muta­ light microscopy is nevertheless recognized as abnormal tion may also have an effect on gene regulation. A pre­ by the cytokinetic machinery. Weak inhibition of topoi- liminary test shows that the prod mutation is not a somerase II activity also results in abnormal anaphases modifier of PEV (data not shown). However, the defects and cell death in Drosophila embryos, whereas stronger of differentiation in mitotic clones of prod and the ab­ inhibition causes visible chromosome undercondensa­ normalities caused by Prod overexpression may be ex­ tion (Buchenau et al. 1993). plained more easily by an effect on gene regulation than The separated sister chromatids in prod mutants are by direct chromosomal effects. We cannot therefore ex­ reminiscent of the PSCS phenotype. In other known mu­ clude the possibility that Prod is also involved in gene tations with a PSCS phenotype the association of chro­ regulation, as are other proteins that influence chromo­ matids with spindles is random, leading to aneuploid some structure. divisions (Kerrebrock et al. 1992; Miyazaki and Orr- Weaver 1992; Rockmill and Roeder 1994), but chromo­ Methods some movement is not affected. However, aneuploid nu­ clei have not been observed in prod chromosome Isolation of deletions by imprecise excision squashes, suggesting that cells with abnormal anaphases The VlacW insert was remobilized by crossing y w; piod/CyO seldom survive until the next division. The lack of an- females to CyO/Sp; Sh^2,3/TM6 males, y W; CyO/prod; euploidy and the scattered arrangement of sister chroma­ SbA.2,3/+ males from the progeny were crossed to yw; Cyo/Sco tids in prod mutants distinguishes the prod phenotype females. White-eyed yw; Cyo/Df{2R)piodl males having lost from the known PSCS phenotype. Sister chromatids of the w* marker of VlacW were backcrossed individually to yw; ACA-injected human cells eventually also disjoin and CyO/prod females. If no viable Cy'^ progeny were found in the display "scattered anaphase" figures (Bernat et al. 1990). next generation because of imprecise excision, a stock was es­ tablished and the line subjected to Southern blot analysis to detect changes in the prod coding region. Euchromatic Prod function Mosaic experiments The reproducible distribution of Prod-positive sites on polytene chromosomes and the undercondensation of Somatic mosaic experiments were carried out using the Minute chromosome arms observed in the mutant indicate that technique, in which the mutant clones are Minute^ in a Minute Prod is also necessary for the condensation of euchroma- heterozygous background and therefore have a growth rate ad­ vantage (Morata and Ripoll 1975). y f^Vy f"'""; f*44 f^52 M(2)iy tin as well as heterochromatin. Assuming that the dis­ CyO homozygous forked mutant females that had two inser­ tribution of Prod on mitotic chromosomes is similar to tions of the forked* gene on 2R, and the M(2)I^ mutation on the that seen on polytene chromosomes, it seems likely that same chromosome arm balanced with CyO (provided by A. Gar- the protein contributes to one of the final levels of eu- cia-Bellido) were crossed to prod/CyO males. Timed egg collec­ chromatin folding (Manuelidis and Chen 1990) by orga­ tions were made at 4 hr intervals and the resulting larvae irra­ nizing megabase-sized pieces of DNA. diated with 1000 R of 7rays at 48 hr. The non-CyO y f ^^VY; Many of the known genes whose products have a dis­ prod/f*44 f*52 M(2)I^ male progeny were mounted in euparal for tinct chromosomal distribution are involved in gene microscopy. Control experiments were done with wild-type Or­ regulation. For example, the products of the Polycomb egon-R chromosomes instead oipiod. Mosaic experiments in a group genes bind to the regulatory sequences of some 60 wild-type (non-Minute) background were done the same way as above, except the 2R forked insertion-bearing chromosome did other genes (Paro 1990). Some modifiers of PEV also bind not carry the Minute mutation. to the chromocenter and many discrete loci on polytene The germ-line mosaic experiment was carried out using the chromosomes (James et al. 1989; Garzino et al. 1992). dominant female sterile technique (Perrimon and Gans 1983). The Trithorax-like [Trl] gene that encodes the GAGA Balanced prod females were crossed to males carrying insertions factor is both an enhancer of PEV and is required for of the dominant female sterile mutation ovoDl at 2R (Chou et normal homeotic gene expression (Farkas et al. 1994). al. 1993). Timed egg collections were made at 4-hr intervals, and The Trl gene is remarkably similar to prod in many re­ the resulting larvae irradiated with 1200 R of ^rays at 70-74 hr spects. Its protein product is localized at the centromeric AEL. ovoDl/pTod virgins were crossed to Oregon-R males in region of Drosophila mitotic chromosomes throughout groups of 15 and checked for progeny. Oregon-R chromosomes the cell cycle (Raff et al. 1994) and many euchromatic were used as controls. loci on polytene chromosomes (Tsukiyama et al. 1994). Null mutants of Trl are late larval lethal with small P-transformation and phenotypic rescue imaginal discs, and weak alleles are semisterile with re­ An 8-kb Hindlll-Xhol fragment from the genomic X clone (Fig.

222 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Molecular and genetic characterization of prod

4) was subcloned into the EcoRl-Xhol site of the PUAST P- whole protein extract from homozygous prod larvae bound to transformation vector (Brand and Perrimon 1993) with a w* Affi-Gel 15. marker. The 1.5-kb Notl-Xhol fragment of the pBluescript(SK)" vector containing the 1563-bp cDNA was also subcloned into the Not-Xho site of the PUAST vector. Plasmid DNA was pu­ Antibody staining rified on Qiagen 500-tip affinity columns (Qiagen, Inc.). P trans­ Embryos were collected at 25°C on apple juice-agar plates, formation was done according to Park and Lim (1995). rinsed in water, and dechorionated for 90 sec in 50% commer­ cial bleach solution. Taxol pretreatment, fixation, and devitel- linization were done as described in Karr and Alberts (1986); Larval neuroblast squash preparations then the embryos were rinsed twice in PBS for 5 min and Preparations were made as described in Protocol 3 of Ashburner blocked in 2% BSA, 0.2% MP40, and 0.02% sodium azide in (1989) without Na citrate treatment and stained with Giemsa. Ix PBS for 2 hr. Purified primary antibody was diluted 1:1000; In the colcemid-treated preparations, ganglia were incubated in anti-(3-tubulin AB (Boehringer Mannheim GmbH) was diluted 0.7% NaCl containing 10 lag/ml of colcemid (Molecular Probes, 1:200 in the above blocking solution and incubated with the Inc.) for 1.5 hr and 0.5% Na citrate for 10 min prior to fixation. embryos at 4°C overnight. The embryos were washed in block­ Mitotic index was determined as the number of cells with vis­ ing solution 4 X 15 min, incubated with fluorescein or Cy3- ible mitotic chromosomes per microscopic field. The field was coupled secondary antibodies (Jackson ImmunoResearch Labo­ defined as the area visible with a lOOx Zeiss objective and lOx ratories Inc.), diluted 1:200 in blocking solution for 2 hr at room eyepieces. For determining the metaphase/anaphase ratio, at temperature, and washed 4x15 min in the same solution. least 300 fields were counted from 30 prod/prod or prodj+ in­ Chromosomes were stained with 1 mM TO-PRO (Molecular dividuals. Probes) in Ix PBS for 2 hr and washed 4x15 min in Ix PBS at room temperature. The embryos were then mounted in Vectashield mounting medium (Vector Laboratories) for confo- Nucleic acid procedures cal microscopy. They were viewed at 60x on a Bio-Rad 1024 laser-scanning confocal microscope. Standard nucleic acid protocols (Sambrook et al. 1989) were Salivary gland polytene chromosomes were prepared as de­ used throughout the study unless otherwise noted. Genomic scribed in Protocol 24 of Ashburner (1989). After removing the DNA fragments next to the VlacW element (Bier et al. 1989) coverslip the slides were immersed in Ix PBS and stained im­ were isolated by plasmid rescue: DNA from heterozygous prod mediately. Chromosome preps were incubated with 50 pi of flies was digested with Bgl\, self-ligated, and transformed into purified Prod antibody diluted 1:200 in PBS, 2% BSA, and 0.2% XL-1 blue-competent cells. The rescued clones were restriction MP40 for 1 hr under coverslip and washed in IxPBS, 0.2% mapped and used as a probe for screening a A. EMBL4 genomic MP40, for 3 X 5 min. Slides were then incubated with 50 pi library and two cDNA libraries made in Lambda ZAPII vectors. fluorescent secondary antibody, diluted 1:200 in the above so­ Bluescript vectors containing cDNA clones were excised from lution for 30 min at room temperature, and washed again for Lambda ZAPII phages with the help of Stratagene Lambda 3x5 min. Preparations were stained in 1 pg/ml of 4', 6-diami- ZAPII Vector Kit. Genomic DNA pieces were subcloned from dino-2-phenylindole (DAPI) (Sigma Chemical Co.) or 15 pM X.EMBL4 into Bluescript for sequencing. Both genomic and propidium iodide (Molecular Probes) for 1 min, covered with cDNA clones were subjected to double-stranded PCR sequenc­ Vectashield mounting medium under coverslip, and examined ing with the double-stranded DNA Cycle Sequencing System of with fluorescence microscopy or confocal microscopy. GIBCO BRL using a series of 15- to 18-mer oligonucleotide primers. Total RNA was purified with the TRlzol reagent of GIBCO Electron microscopy BRL. Poly(A)* RNA was affinity purified by the poly(AT) tract mRNA isolation system of Promega, electrophoresed (2 pg/ Lymph glands and imaginal discs of mutant and Oregon-R con­ lane) on 1.0% agarose/0.66 M formaldehyde gels, transferred trol larvae were dissected in PBS and fixed in 2% glutaraldehyde with 20x SSC to positively charged nylon membrane (Amer- 1X PBS. The tissue was washed several times in 0.1 M cacodylate sham), UV cross-linked, and hybridized at 42°C in 50% form- buffer containing 8% sucrose and postfixed in osmium tetrox- amide with random primed '^^P-labeled probe (Stratagene Prime- ide, stained with lead citrate and uranyl acetate, embedded in It kit). plastic, and sectioned for transmission electron microscopy.

Fusion protein and antibody production Acknowledgments The 1526-bp £coRI fragment containing the entire coding region We thank members of the Bryant laboratory, especially Dan from the 1563-bp cDNA clone was subcloned into the £coRI Woods, Harald Biessmann, Marika Walter, and Mike Boedighei- site of pGEX-4T-2 vector (Pharmacia). Cells carrying this con­ mer for technical advice and fruitful discussions. The Gal4 lines struct were induced, harvested, and lysed according to the com­ were provided by N. Perrimon. A 0- to 24-hr embryonic library pany's protocol. Fusion protein was purified from inclusion bod­ was provided by Carl S. Thummel, and a third-instar larval li­ ies by two rounds of PAGE and sent for antibody production to brary was given by Rudi Grams. A chromosome carrying two Pocono Rabbit Farm and Laboratory, Inc. (Canadensis, PA). insertions of the forked* gene on 2R and the M(2)f mutation on The 850-bp BamHl-Xhol fragment containing the 3' half of the same chromosome arm was provided by Antonio Garcia- the cDNA was also subcloned into pGEX-4T-2, induced, har­ Bellido. This investigation was supported by a Long-term Fel­ vested, and fusion protein purified as described above. Antise­ lowship from the International Human Frontier Science Pro­ rum was affinity purified with the pGEX-850-bp fusion protein, gram (HFSPO) to T.T. (no. LT-213/94) and by grant no. using Affi-Gel 15 (Bio-Rad) and following the company's proto­ HD02713 from the National Institutes of Health. The nucleo­ col. Remaining background activity was removed from the af­ tide sequence of prod cDNA has been deposited in Genbank finity-purified antibody with a subtraction column containing under Accession number U83596.

GENES & DEVELOPMENT 223 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Torok et al.

The publication costs of this article were defrayed in part by proteins in the chromosomal events of mitosis. FASEB J. payment of page charges. This article must therefore be hereby 8: 947-956. marked "advertisement" in accordance with 18 USC section Ellison, J.R. and G.C. Howard. 1981. Non-random position of 1734 solely to indicate this fact. the A-T rich DNA sequences in early embryos of Drosophila virilis. Chromosoma 83: 555-561. Farkas, G., J. Gausz, M. Galloni, G. Renter, H. Gyurkovics, and References F. Karch. 1994. The Trithorax-like gene encodes the Dro­ sophila GAGA factor. Nature 371: 806-808. AUshire, R.C. 1995. Elements of chromosome structure and Fischer, J.A., E. Giniger, T. Maniatis, and M. Ptashne. 1988. function in fission yeast. Semin. Cell Biol. 6: 55-64. GAL4 activates transcription in Drosophila. Nature AUshire, R.C, E.R. Nimmo, K. Ekwall, J.P. Javerzat, and G. 332: 853-856. Cranston. 1995. Mutations derepressing silent centromeric Foe, V.E. 1989. Mitotic domains reveal early commitment of domains in fission yeast disrupt chromosome segregation. cells in Drosophila embryos. Development 107: 1-22. Genes & Dev. 9: 218-233. Foe, V.E. and B.M. Alberts. 1985. Reversible chromosome con­ Ashburner, M. 1989. Drosophila: A laboratory handbook. Cold densation induced in Drosophila embryos by anoxia: Visu­ Spring Harbor Laboratory Press, Cold Spring Harbor, NY. alization of interphase nuclear organization. /. Cell Biol. Baum, M., V.K. Ngan, and L. Clarke. 1994. The centromeric 100: 1623-1636. K-type repeat and the central core are together sufficient to Gall, J.G., E.H. Cohen, and M.L. Polan. 1971. Repetitive DNA establish a functional Schizosaccharomyces pombe centro­ sequences in Drosophila. Chromosoma 33: 319-344. mere. Mol. Biol. Cell 5: 747-761. Garzino, V., A. Pereira, P. Laurenti, Y. Graba, R.W. Levis, Y. Le Bernat, R.L., G.G. Borisy, N.F. Rothfield, and W.C. Earnshaw. Parco, and J. Pradel. 1992. Cell lineage-specific expression of 1990. Injection of anticentromere antibodies in interphase modulo, a dose-dependent modifier of variegation in Dro­ disrupts events required for chromosome movement at mi­ sophila. EMBO f. 11: 4471-4479. tosis, f. Cell Biol. Ill: 1519-1533. Gatti, M. and B.S. Baker. 1989. Genes controlling essential cell- Bernat, R.L., M.R. Delannoy, N.F. Rothfield, and W.C. Earn­ cycle functions in Drosophila melanogaster. Genes Si Dev. shaw. 1991. Disruption of centromere assembly during in­ 3:438-453. terphase inhibits kinetochore morphogenesis and function Gatti, M. and M.L. Goldberg. 1991. Mutations affecting cell in mitosis. Cell 66: 1229-1238. division in Drosophila. Methods Cell Biol. 35: 543-586. Bhat, K.M., G. Farkas, F. Karch, H. Gyurkovics, J. Gausz, and P. Gatti, M., S. Pimpinelli, and G. Santini. 1976. Characterization Schedl. 1996. The GAGA factor is required in the early Dro­ of Drosophila heterochromatin. I. Staining and decondensa­ sophila embryo not only for transcriptional regulation but tion with Hoechst 33258 and quinacrine. Chromosoma also for nuclear division. Development 122: 1113-1124. 57:351-375. Bier, E., H. Vaessin, S. Shepherd, K. Lee, K. McCall, S. Barbel, L. Glover, D.M. 1989. Mitosis in Drosophila. /. Cell Sci. 92: 137- Ackerman, R. Carretto, T. Uemura, E. Grell, et al. 1989. 146. Searching for pattern and mutation in the Drosophila ge­ James, T.C., J.C. Eissenberg, C. Craig, V. Dietrich, A. Hobson, nome with a P-lacZ vector. Genes & Dev. 3: 1273-1287. and S.C. Elgin. 1989. Distribution patterns of HPl, a hetero- Brand, A.H. and N. Perrimon. 1993. Targeted gene expression as chromatin-associated nonhistone chromosomal protein of a means of altering cell fates and generating dominant phe- Drosophila. Eur. J. Cell Biol. 50: 170-180. notypes. Development 118:401-415. Kapuscinski, J. 1995. DAPI: A DNA-specific fluorescent probe. Buchenau, P., H. Saumweber, and D.J. Arndt-Jovin. 1993. Con­ Biotechnal. Histochem. 70: 220-233. sequences of topoisomerase II inhibition in early embryogen- Karpen, G.H., M.-H. Le, and H. Le. 1996. Centric heterochro­ esis of Drosophila revealed by in vivo confocal laser scanning matin and the efficiency of achiasmate disjunction in Dro­ microscopy. /. Cell Sci. 104: 1175-1185. sophila female meiosis. Science 273: 118-122. Carmena, M., J.P. Abad, A. Villasante, and C. Gonzalez. 1993. Karr, T.L. and B.M. Alberts. 1986. Organization of the cytoskel- The Drosophila melanogaster dodecasatellite sequence is eton in early Drosophila embryos. /. Cell. Biol. 102: 1494- closely linked to the centromere and can form connections 1509. between sister chromatids during mitosis. /. Cell Sci. Kerrebrock, A.W., W.Y. Miyazaki, D. Birnby, and T.L. Orr- 105:41-50. Weaver. 1992. The Drosophila mei-S332 gene promotes sis- Cavener, D.R. 1987. Comparison of the consensus sequence ter-chromatid cohesion in meiosis following kinetochore flanking translational start sites in Drosophila and verte­ differentiation. Genetics 130: 827-841. brates. Nucleic Acids Res. 15: 1353-1361. Li, X. and R.B. Nicklas. 1995. Mitotic forces control a cell-cycle Chou, T.B., E. Noll, and N. Perrimon. 1993. Autosomal P[o- checkpoint. Nature 373: 630-632. voDl] dominant female-sterile insertions in Drosophila and Lobe, A.R., A.J. Hilliker, and P.A. Roberts. 1993. Mapping their use in generating germ-line chimeras. Development simple repeated DNA sequences in heterochromatin of Dro­ 119:1359-1369. sophila melanogaster. Genetics 134: 1149-1174. Clarke, L. 1990. Centromeres of budding and fission yeasts. Manuelidis, L. and T.L. Chen. 1990. A unified model of eukary- Trends Genet. 6: 150-154. otic chromosomes. Cytometry 11: 8-25. DiNardo, S., K. Voelkel, and R. Stemglanz. 1984. DNA topoi­ Miklos, G.L. and J.N. Cotsell. 1990. Chromosome structure at somerase II mutant of Saccharomyces cerevisiae: Topoisom­ interfaces between major chromatin types: alpha- and beta- erase II is required for segregation of daughter molecules at heterochromatin. BioEssays 12: 1-6. the termination of DNA replication. Proc. Natl. Acad. Sci. Miyazaki, W.Y. and T.L. Orr-Weaver. 1992. Sister-chromatid 81:2616-2620. misbehavior in Drosophila ord mutants. Genetics Doane, W.W. and F.E. Dumapias. 1987. Deletion limits of the 132: 1047-1061. adp, map and nw loci in Drosophila melanogaster. Dros. Inf. . 1994. Sister-chromatid cohesion in mitosis and meiosis. Service 66: 49. Annu. Rev. Genet. 28: 167-187. Earnshaw, W.C. and A.M. Mackay. 1994. Role of nonhistone Morata, G. and P. RipoU. 1975. Minutes: Mutants of Drosophila

224 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Molecular and genetic characterization of prod

autonomously affecting cell division rate. Dev. Biol. Biol. 111:2839-2850. 42:211-221. Woods, D.F., C. Hough, D.J. Peel, G. Callaini, and P.J. Bryant. Murphy, T.D. and G.H. Karpen. 1995. Localization of centro­ 1996. Dig protein is required for junction structure, cell po­ mere function in a Drosophila minichromosome. Cell larity, and proliferation control in Drosophila epithelia. /. 82:599-609. Cell Biol. 134: 1469-1482. O'Brien, T., R.C. Wilkins, C. Giardina, and J.T. Lis. 1995. Dis­ Wu, C.I., T.W. Lyttle, M.L. Wu, and G.F. Lin. 1988. Association tribution of GAGA protein on Drosophila genes in vivo. between a satellite DNA sequence and the Responder of Seg­ Genes & Dev. 9: 1098-1110. regation Distorter in D. melanogaster. Cell 54: 179-189. Pardue, M.L. and W. Hennig. 1990. Heterochromatin: Junk or collectors item? Chromosoma 100: 3-7. Park, S. and J.K. Lim. 1995. A microinjection technique for etha- nol-treated eggs and a mating scheme for detection of germ line transformants. Dros. Inf. Service 76: 187-189. Paro, R. 1990. Imprinting a determined state into the chromatin of Drosophila. Trends Genet. 6: 416-421, Perrimon, N. and M. Gans. 1983. Clonal analysis of the tissue specificity of recessive female-sterile mutations of Dro­ sophila melanogaster using a dominant female-sterile muta­ tion Fs(l)K1237. Dev. Biol. 100:365-373. Pluta, A.F., A.M. Mackay, A.M. Ainsztein, I.G. Goldberg, and W.C. Eamshaw. 1995. The centromere: Hub of chromosom­ al activities. Science 270: 1591-1594. Radic, M.Z., K. Lundgren, and B.A. Hamkalo. 1987. Curvature of mouse satellite DNA and condensation of heterochroma­ tin. Cell 50: 1101-1108. Raff, J.W., R. Kellum, and B. Alberts. 1994. The Drosophila GAGA transcription factor is associated with specific re­ gions of heterochromatin throughout the cell cycle. EMBO J. 13:5977-5983. Rockmill, B. and G.S. Roeder. 1994. The yeast medl mutant undergoes both meiotic homolog nondisjunction and preco­ cious separation of sister chromatids. Genetics 136: 65-74. Saka, Y., T. Sutani, Y. Yamashita, S. Saitoh, M. Takeuchi, Y. Nakaseko, and M. Yanagida. 1994. Fission yeast cut3 and cut 14, members of a ubiquitous protein family, are required for chromosome condensation and segregation in mitosis. EMBO J. 13: 4938-4952. Sambrook J., E.F. Fritsch, and T. Maniatis. 1989. Molecular cloning: A laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Shrestha, R. and E. Gateff. 1982. Ultrastructure and cytochem­ istry of the cell types in the larval hematopoietic organs and hemolymph of Drosophila melanogaster. Dev. Growth Dif­ fer. 24: 65-82. Sluder, G. 1990. Functional properties of kinetochores in animal cells. Curr. Opin. Cell Biol. 2: 23-27. Torok, T., G. Tick, M. Alvarado, and I. Kiss. 1993. P-lacW in- sertional mutagenesis on the second chromosome of Dro­ sophila melanogaster: Isolation of lethals with different overgrowth phenotypes. Genetics 135: 71-80. Tsukiyama, T., P.B. Becker, and C. Wu. 1994. ATP-dependent nucleosome disruption at a heat-shock promoter mediated by binding of GAGA transcription factor. Nature 367: 525- 532. Willard, H.F. 1990. Centromeres of mammalian chromosomes. Trends Genet. 6: 410-416. Wilson, C, R.K. Pearson, H.J. Bellen, C.J. O'Kane, U. Grossniklaus, and W.J. Gehring. 1989. P-element-mediated enhancer detection: An efficient method for isolating and characterizing developmentally regulated genes in Dro­ sophila. Genes & Dev. 3: 1301-1313. Wines, D.R. and S. Henikoff. 1992. Somatic instability of a Dro­ sophila chromosome. Genetics 131: 683-691. Wood, E.R. and W.C. Earnshaw. 1990. Mitotic chromatin con­ densation in vitro using somatic cell extracts and nuclei with variable levels of endogenous topoisomerase II. /. Cell

GENES & DEVELOPMENT 225 Downloaded from genesdev.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

The product of proliferation disrupter is concentrated at centromeres and required for mitotic chromosome condensation and cell proliferation in Drosophila.

T Török, P D Harvie, M Buratovich, et al.

Genes Dev. 1997, 11: Access the most recent version at doi:10.1101/gad.11.2.213

References This article cites 60 articles, 27 of which can be accessed free at: http://genesdev.cshlp.org/content/11/2/213.full.html#ref-list-1

License

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the top Service right corner of the article or click here.

Copyright © Cold Spring Harbor Laboratory Press